- AWS announced updates and services that promise to boost performance and lower compute costs Tuesday, just weeks after becoming a premier member of the FinOps Foundation, a nonprofit trade group developing cloud billing and optimization standards.
- The company unveiled Graviton4, the next generation of its cost- and performance-optimized CPU technology. The chips will “provide the best price performance and energy efficiency for memory-intensive workloads,” CEO Adam Selipsky said during his Tuesday keynote at AWS re:Invent 2023.
- The cloud provider also extended a strategic partnership with Nvidia to integrate the chipmaker's new H200 processors into AWS infrastructure and unveiled Trainium2, the second generation of the hyperscaler’s lower-cost alternative for processing LLM and ML workloads.
While global cloud spending continues to rise, AWS and its fellow hyperscalers have been trying to keep customers happy by helping them control cost.
More compute for less money, a central tenet of cloud, has been harder to achieve than anticipated for many organizations. As enticing yet costly SaaS features multiply and post-migration usage creeps up, IT is left holding the bill.
AWS pledged to help stem cloud overspend last year, when economic sentiment soured, and has continued to prioritize workload optimization assistance.
Faster and more cost-efficient data center processing power built on a new generation of chips is part of a broader spectrum of remedies Selipsky highlighted.
To optimize analytics operations, AWS introduced a new storage class, S3 Express One Zone, which uses purpose-built hardware and software to accelerate data processing. AWS is promising up to 60% savings on compute costs and data access at half the price of the S3 Standard tier.
The company also announced a preview of Amazon Redshift Serverless, an AI-enabled cloud data warehousing solution that automates scaling across data volume changes, concurrent users and query complexity. Internal tests yielded an up to tenfold improvement in price performance, according to AWS.
One of the day’s biggest reveals, multifunctional generative AI-powered enterprise assistant Q, has optimization potential, too, Selipsky said.
Trained on the detailed minutiae of AWS’ infrastructure and services, the natural language chatbot can provide migration assistance and direct users to right-sized storage and compute.
“You can just tell Q about your workload and get an accurate, quick and economical instance recommendation,” Selipsky said.